This site collects resources related to implementing Archival Repository (RRKive) solutions as part of a Data Commons based on a set of Protocols know as the PILARS: Principles for Implementing Long-term Archival Repository Services.

The PILARS are maintained as stand-alone project managed in a Git repository.

What’s a Data Commons and where’s the data in one?

Following the lead of Jenny Fewster, HASS and Indigenous Research Data Commons Director at the Australian Research Data Commons (ARDC) we use this definition for Data Commons (Grossman et al. (2016)):

A global trusted system of systems that provides frictionless access to high quality interoperable resources, services and artefacts for research.

The abstract of this article says:

Data commons collocate data, storage, and computing infrastructure with core services and commonly used tools and applications for managing, analyzing, and sharing data to create an interoperable resource for the research community.

NOTE: This definition differs from others in that it does not use the word “open”, which is important, as not all research data can be made openly available; according to the CARE and FAIR principles can and should be made accessible to the right people / agents. For some background on CARE and FAIR in a Data Commons context see this Language Data Commons of Australia blog post.

This web site looks at the core services and infrastructure needed to undertake the processes mentioned in the definition: managing, analyzing, and sharing; we do that below, paying particular attention to where the data resides in a Data Commons.

This is about Research Data Commons deployments

This site is about:

  1. Implementation advice for the PILARS protocols.

  2. A toolkit for deploying granular sustainable archival repository software which can describe and make data accessible down to and inside of the files and datasets but using commodity IT systems to ensure data interoperability

  3. Enabling the ‘interoperability’ mentioned in the definition of a Data Commons above.

One of the key inspirations for the PILARS was the approach taken by PARADISEC where data is stored and managed using a very simple architecture with data and metadata at its heart sitting, on top of commodity IT services; initially this was a file-system and is now cloud-based object storage. The key idea was that the data should always be available to administrators independently of particular software services (though because some data is not open it is not possible to just put data up on an open webserver – access control is needed).

This site is an evolution of a previous effort know as Arkisto, see the background page.

Scope: what people, domains, institutions, kinds and scales of data is this relevant for?

This site is for leaders and implementers of Research Data Commons projects AND general research data management practitioners looking to choose, manage or establish sustainable CARE and FAIR compliant data management solutions that will work with research services.

This site is relevant to any research or cultural domain, where teams are establishing data management infrastructure – the initial uses case and implementations are mainly from the Humanities and Social Sciences, with some “sciencey” and generalist deployments.

Not in scope

There are a number of data management / Data Commons scenarios where this will be less relevant; where very bit of existing and prospective data, and all analytical products and annotations have an identified home with:

  • Persistent Identification
  • Sustainable secure storage for raw and derived data and research outputs at useful granularity – eg down to individually addressable items, files or variables within files
  • Appropriate access control
  • Catalogues / portals to make data discoverable
  • APIs to integrate and interoperate with analytical and data curation processes

The website is an initiative of the Language Data Commons of Australia; one of the partners of the Australian Research Data Commons (ARDC).